Acceptable Use Policy

Effective Date: September 28, 2026

This Acceptable Use Policy (“AUP”) applies to all Deyra products, Subscription Services, APIs, integrations, and related functionality provided by Five Jars Corp. It forms part of the Agreement under the Deyra Terms & Conditions. Capitalized terms not defined here have the meanings in those Terms.

The general rules below apply to Customer, Authorized Users, and End Users across the Platform. Additional AI restrictions apply only when AI Features are used. Customer must take reasonable steps to prevent and address misuse of its implementation and comply with the deployment obligations in the Terms.

1. General Rules for All Services

Illegal Activity, Harm, and Infringement

You may not use the Platform to:

• Engage in or facilitate unlawful activity, fraud, scams, money laundering, market manipulation, or violations of applicable export controls or sanctions;
• Infringe intellectual-property, privacy, publicity, or other rights, including unauthorized use of copyrighted works, confidential information, or trade secrets;
• Facilitate violence, terrorism, violent extremism, human trafficking, exploitation, or abuse;
• Create, obtain, distribute, or facilitate child sexual abuse material, sexualization or exploitation of children, grooming, or inappropriate contact with minors;
• Harass, threaten, stalk, doxx, defame, or unlawfully discriminate against individuals or groups, or promote hatred or violence based on protected characteristics;
• Distribute non-consensual intimate imagery, or sexually explicit or pornographic content outside a lawful research context expressly approved by Five Jars;
• Promote self-harm, suicide, eating disorders, or dangerous conduct likely to cause serious injury;
• Facilitate unlawful gambling or unlawful sale or distribution of weapons, controlled substances, or other regulated goods.

Security, Access, and Platform Abuse

You may not:

• Access accounts, systems, data, or networks without authorization, or introduce malware or malicious code;
• Probe, scan, or test Platform vulnerabilities without Five Jars’ prior authorization;
• Circumvent security controls, permissions, technical safeguards, rate limits, or contracted usage limits, including through multiple accounts;
• Launch denial-of-service attacks, spam campaigns, or other activity that disrupts the Platform or other customers;
• Impose unreasonable load or intentionally degrade service performance;
• Share credentials or allow unauthorized access to an account;
• Reverse engineer, decompile, or extract Platform source code or models except where applicable law expressly permits it notwithstanding this restriction;
• Scrape, harvest, or extract data where you lack authorization or a lawful right to do so.

Usage and capacity limits follow the applicable Order Form and the Terms. A reference to unlimited usage does not permit abusive traffic or circumvention of technical safeguards.

Privacy, Data Rights, and Sensitive Information

You may not:

• Submit, connect, disclose, or process information without the necessary rights, lawful basis, notices, and consents;
• Use the Services for unauthorized surveillance, tracking, or identification of individuals;
• Use the Services to identify individuals from images or biometric data;
• Collect or process sensitive personal information, including health, biometric, or genetic information, without the legal basis, safeguards, and any separate agreements required for that use;
• Submit information subject to a legal requirement for a separate provider agreement, such as a business associate agreement, before that agreement has been executed where required;
• Expose private information through an improperly configured public interface or attempt to obtain another customer’s data.

Impersonation, Deception, and Manipulation

You may not impersonate a person, organization, public official, or Five Jars representative; misrepresent affiliations; create fraudulent credentials or documents; run phishing or scam campaigns; or manipulate reviews, testimonials, or public communications through deceptive content. You may not use the Services for voter suppression, election interference, false voting instructions, or intentionally deceptive disinformation campaigns.

2. Minors and Customer-Facing Deployments

Business account holders and administrators must be at least 18. Customer-facing Services may be used by minors only with age-appropriate safeguards and compliance with applicable law. Customer must provide appropriate notices, supervision, and parental consent mechanisms where legally required.

A deployment directed to children under 13, or involving knowing online collection of personal information from children under 13, requires Five Jars’ prior written approval and documented data-processing and consent arrangements. Requirements under applicable law protecting children at other ages also apply. Approval does not replace required notice, verifiable parental consent, or any other legal obligation. The Children’s Privacy section describes the parties’ roles and the reporting route for concerns. Five Jars remains responsible for legal obligations that apply directly to it.

Child exploitation and other prohibited harmful conduct cannot be authorized by parental consent, Customer instructions, or Five Jars’ approval.

3. Additional Restrictions for AI Features

This section applies only to functionality that uses AI Features. The general rules above continue to apply to AI inputs, outputs, and deployment.

Harmful Generation and Deceptive AI

You may not use AI Features to:

• Generate instructions that facilitate violence, terrorism, weapons or explosives manufacture, illegal drug manufacture, or other conduct likely to cause serious harm;
• Create exploitative, hateful, discriminatory, or other prohibited content described in Section 1;
• Generate deceptive deepfakes, synthetic identities, fake credentials, or impersonations intended to mislead people about identity, affiliation, authority, or authenticity;
• Create targeted political disinformation, false voting information, deceptive election predictions presented as official results, or undisclosed synthetic political communications intended to deceive;
• Fabricate authoritative medical, legal, financial, public-health, or emergency information likely to cause harm;
• Manipulate religious or cultural content, impersonate religious authorities, or falsely claim authority to incite hatred, violence, or exploitation;
• Misrepresent AI outputs as human-generated or human-verified when they are not, or make false or misleading claims about AI capabilities;
• Circumvent model safeguards, conduct unauthorized prompt-injection attacks, or extract hidden instructions, credentials, protected content, or other customers’ information;
• Use the Services or their outputs to develop competing AI products or services.

Authorized security testing remains subject to its expressly agreed scope. No approval permits illegal or exploitative uses.

Professional Advice and Consequential Decisions

Do not use AI Features to provide medical diagnosis or treatment advice, legal advice or legally binding commitments, investment or financial advice, or other regulated professional services without Five Jars’ prior written approval, documented independent legal review, and appropriate review and oversight by qualified professionals.

Do not make employment, credit, insurance, eligibility, or other decisions materially affecting a person’s rights or access to essential opportunities solely on AI outputs. Any proposed AI deployment for those decisions requires Five Jars’ prior written approval, documented independent legal review, and meaningful human oversight and safeguards. Approval does not establish that a use is lawful or remove professional-licensing requirements.

High-Risk Deployments Requiring Prior Approval

In addition to any other applicable requirements, obtain Five Jars’ prior written approval and document independent legal review before deploying AI Features in:

• Clinical, financial, or legal workflows where errors could cause patient harm, significant financial loss, or harm to legal rights;
• Safety-critical infrastructure, transportation, utilities, emergency response, or life-safety systems;
• Government or military applications;
• Child-directed educational experiences requiring enhanced safeguarding, including the under-13 deployments described in Section 2;
• Any context where AI errors could cause death, serious bodily injury, or significant property damage.

A standard subscription does not authorize these deployments. Contact [email protected] with the proposed use, data categories, audience, and safeguards before deployment if the classification is uncertain. Written approval may impose additional controls, contractual terms, or use limitations and cannot waive applicable law or the absolute prohibitions in this AUP.

Ordinary factual display of schedules, activities, prices, or general educational and wellness information does not by itself constitute regulated professional advice or a consequential decision. Where AI is used to generate or interpret such information, Customer must provide suitable limitations and verification guidance, avoid personalized professional advice outside an approved deployment, and apply any relevant minors safeguards. General wellness content should direct users to qualified medical professionals for individual health decisions.

AI Disclosure and Oversight

When End Users interact with AI Features, Customer must clearly disclose the AI interaction before or at first interaction, explain that outputs can contain errors, provide an appropriate verification or human-contact route, and comply with applicable disclosure law. Disclosures must be prominent and appropriate to the audience. Customer must test and monitor its implementation and report harmful or systematic issues as described in the Additional Terms for AI Features. Human review does not substitute for approval where approval is required.

4. Monitoring and Enforcement

Five Jars may investigate reported or detected abuse, monitor operational usage patterns, apply abuse-detection safeguards, and review relevant Customer Data or logs as reasonably necessary to investigate a specific issue. Such access is limited by applicable law, the Agreement, and any executed DPA; it is not authorization to use Customer content for general model training or unrelated purposes.

Enforcement may include a warning, required corrective action, restriction of affected functionality, suspension, or termination under the Terms’ notice and cure provisions. Immediate action may be necessary for unlawful activity, active security threats, material abuse, or legal requirements. We consider the severity, scope, and recurrence of a violation and restore access when the grounds for suspension are resolved, as provided in the Terms.

Refunds, remaining committed fees, and termination consequences are governed by the Terms and applicable Order Form; this AUP does not create a separate blanket forfeiture rule. We may seek lawful remedies or report matters to competent authorities where required or permitted by law.

5. Reporting Violations

Send reports to [email protected] with the subject “AUP Violation Report.” Include the affected Service or Customer implementation, a description, relevant times, and available evidence. Do not include credentials, unnecessary personal information, or copies of illegal material. Reports are reviewed under our internal procedures; we may be unable to disclose enforcement details.

6. Updates to This AUP

We may update this AUP to address new risks, clarify requirements, or comply with law. Material changes will be communicated by email or platform notification at least 30 days before taking effect, except where an earlier change is reasonably required by law or to address an active security or abuse threat. Changes remain subject to the Terms’ protections for accepted Order Forms and any executed DPA. An AUP update does not change agreed pricing or authorize new data uses.

7. Contact

• Email: [email protected]
• Website: deyra.ai